sql row constructor...works!

From: Merlin Moncure <mmoncure(at)gmail(dot)com>
To: pgsql-hackers(at)postgresql(dot)org
Subject: sql row constructor...works!
Date: 2006-02-07 17:31:05
Message-ID: b42b73150602070931o73486e27v2244a91a193bd6e5@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Thanks for bringing the SQL 92 row constructor into spec for
operations involving > and <. This is just fantastic. I just
benchmarked ISAM style access to tables on multi-part keys and the
speedup is tremendous vs. the non row-constructor approach to the
problem which is tedius to write and only selective to the first part
of the key. We are talking speedups proportional to the inselectivity
of the ordering key. While this feature is relatively esoteric, it is
powerful and belongs in every dba's bag of tricks.

I was not expecting full indexabilty on the operation but this too
works properly and efficiently. Great work! IMO, most people are not
aware of a properly working row constructor because it is either not
implemented or improperly implemented. Once again, the team has shown
that it can meet or defeat the big name databases on the feature
level.

Merlin

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Josh Berkus 2006-02-07 17:33:57 Re: sql row constructor...works!
Previous Message Csaba Nagy 2006-02-07 17:17:02 Re: streamlined standby procedure